For years people like me have carped that not only is the public cloud a small slice of IT spending, but the big public clouds are also much more likely to attract startups or other digital natives as customers instead of conventional enterprises.
That begs an important question: What then exactly is the “enterprise cloud”? Recently I interviewed Bob Weiler, executive vice president of global business units at Oracle, to drill into that topic. As head of Oracle Industries, it’s his job to marshal Oracle’s full portfolio of technologies to serve seven key verticals — and Oracle’s cloud is a big part of that push.
When Oracle says “cloud,” one of the first things to remember is that most of the company’s cloud revenue comes from SaaS, despite Larry Ellison’s big talk about going head-to-head against AWS in full-on IaaS combat. SaaS makes particular sense in serving vertical industries. Through acquisition and internal development, Oracle has acquired the expertise to serve, say, the utilities or pharmaceutical or financial services industries through vertical SaaS applications.
Weiler says that Oracle’s most successful acquisitions have been vertical vendors that began life as SaaS plays:
Again, this is a very different play than what AWS, Azure, or Google Cloud do. They’re essentially multitenanted self-service offerings; arguably, multitenancy and self-service define the cloud. Oracle, on the other hand, leans toward a have-it-your-way approach. Whether SaaS, PaaS, or IaaS, Oracle offers not only multitenancy but also single-tenancy and managed cloud services across its entire portfolio — as well as, in many cases, on-premises versions. Rather than click choices on an AWS menu to fire up a new project, you call Oracle sales and work out your special deal.
As a sales and marketing organization, that’s what Oracle does. But that’s also the nature of the enterprise cloud: Regulatory constraints and sheer complexity loom large and demand special treatment. As Weiler says, “If I have an application that is a deposit application at a bank, that’s not going in a multitenancy cloud, or we have a problem — but maybe an analytics application or maybe a lending or leasing application or maybe a customer origination application, well, that’s OK.”
Weiler cites other obstacles. Many organizations cling to legacy applications they developed themselves, even though the proprietary advantage of running them versus configurable vertical SaaS applications may be nonexistent. More importantly, a big swath of companies need to rationalize and refactor not only their applications, but also their business processes on prem first. There’s no sense in moving a big mess into the cloud.
In the tech industry, we like to entertain the notion that enterprises moving to the cloud wholesale is inevitable — just look at the successful examples! But when you do, you find that many of those customers are internet companies or separate operations within enterprises, not the whole hairball of business logic and core systems and mysterious boxes that keep chugging away long after institutional memory of how they work has faded. Untangling that mess takes time.
That’s not to say that the enterprise cloud will ultimately center on enterprise SaaS applications offered by Oracle or SAP or other incumbent players. Or rather, they may adopt those SaaS applications for core business processes that change infrequently, but turn to cloud-native platforms offered by AWS, Azure, or Google Cloud to build dynamic applications that demand continuous change. The IaaS leaders have already gone a long way toward offering every service a developer could want, from machine learning to event processing to new serverless programming paradigms.
Either way, it’s a major mistake to underestimate the magnitude of the enterprise migration to the cloud. “This is a long journey,” says Weiler. “It’s going to take years and years for people to get 100 percent over to the cloud. I don’t think two or three years. This is going to be a 10-year journey.”